The acceleration of heavy nuclei in solar flares

E%ww#. Sjkwe Scl.. Vol. 23. PP. 955 to 959.

Perk%umm Press, 1975.

THE ACCELERATION

Printed in Northern J&and

OF HEAVY NUCLEI

IN SOLAR FLARES

K~TOMO SAKURAI” Radio Astronomy Branch, Laboratory for Bxtraterrestrial Physics NASA, Goddard Space Flight Center, Greenbelt, Md 20771, U.S.A. (Received infinn form 30 October 1974)

A&tract--The overabundance of heavy nuclei in solar cosmic rays of energy g 10 MeV]nucleon (sometimes up to -30 MeV/nucleon) is explained by taking into account the pre-flare ionization states of these nuclei in the region where they are accelerated. A model is proposed which considers two-step accelerations associated with the initial deveIopment of solar flares. The first step is closely related to the triggering process of flares, while the second one starts with the development of the explosive phase. Further ionization of medium and heavy nuclei occnrs through their interaction with keV electrons accelerated by the first-step acceleration. It is suggested that the role of these electrons is important in producing fully ionized atoms in the acceleration regions. 1. INTRODUCTION It is known that the measurement of the chemical composition of solar cosmic rays is very important in the understanding of the acceleration mechanism of these particles and the chemical composition of the solar atmosphere. In fact, the study of the acceleration mechanism must consider the chemical composition of solar cosmic rays in reference to that of the Sun because the comparison between these two compositions would give us an information as to bow solar cosmic rays are accelerated and then ejected into outer space (e.g. Sakurai, 1974a,b). With the extension of the observable lower limit of cosmic ray energy because of the improvement of observing techniques on-board satellites, the discrepancy between the chemical abundances of solar cosmic rays and those of the solar atmosphere has become known in regard to the overabundance of the heavy rmclei in solar cosmic rays of energy GlO MeV/nucleon, sometimes up to ~30 lvIeV/ nucleon (see, e.g. Price et al., 1971; Cowsik and Price, 1971; Armstrong and Krimigis, 1971; Armstrong et at., 1972; MogroCampero and Simpson, 1972a). However it is also known that, for solar cosmic rays of energy 210 MeV~~ucleon, their nuclear abundances, in general, are very similar to those of the solar atmosphere (e.g. Biswas and Fichtel, 1964, 1965; Bertsch et al., 1969, 1972, 1973; Price, 1973). This difference mentioned above may be related to the acceleration processes of solar cosmic rays in or near the flare region (MogroCamper0 and Simpson, 1972a,b; Caught and *NASA Associate with University Astronomy Program.

of Maryland,

Mogro-Campero, 1972, 1973). The recent observational results on Solar cosmic ray compositions are summarized in Hovestadt (1973). At present, on the development of solar flares, we have a lot of observational materials regarding particle and electromagnetic radiations. They are considered in est~ating the real processes of solar flares. By taking into account the development pattern of solar flares, we will here consider the acceleration of the heavy nuclei of solar cosmic rays in solar flares. In particular, discussion will be given on the origin of the overab~dan~ of the heavy nuclei in solar cosmic rays in the low energy range ( G 10 MeV/nudeon). 2. IONIZATION STATRS OF THR HEAVY NU~IIN&REFLARE It seems likely that solar cosmic ray nuclei and relativistic electrons are s~~t~eousIy accelerated in a different region from that in which the Har emissions are generated (e.g. de Jager, 1969; Svestka, 1966, 1969; Sakurai, 197la). It is estimated that the former region, hereafter called the “acceleration region,” is located higher up in the solar atmosphere, the density of which is 10s to lOr* particles per cma (Sakurai, 1971a; Kane and Lin, 1971). The temperature of this region seems to be equal to that of the medium surrounding this region, say, I@-lo6 K. Furthermore, the intensity of the ambient sunspot magnetic fields is estimated to be 10-1OOG above the active sunspot groups (Sakurai, 197la). Since this region seems to be thermally in quasi-equilibrium before the onset of a flare, the ionization states of the iron-group 955

particles are estimated by using the results calculated by Jordan (1969,197O). The result for the relative abundances of the iron atoms in different ionization states is shown in Fig. 1. This figure shows that most of the iron atoms are not fully ionized, but are in the charge states from FeIX to FeXVII in the medium of a temperature 106-4 x 106K. If this temperature is reduced to -2 x 105K, for example, the ionization states are highIy reduced, to FeVI to FeVII. A similar situation also occurs in the Ni atoms for the temperatures as mentioned above (e.g. Jordan, 1970). Even the oxygen and silicon atoms are not fully ionized in the medium of these temperatures. The results obtained here, therefore, suggest that there are a number of heavy ions not fully ionized in the acceleration region before a flare starts. These ions seem to be accelerated to high energy in association with the development of the explosive phase of a flare, though some of them seem to become more ionized through various processes accompanied by the acceleration. Therefore, it seems unlikely that all of the heavy ions are fully ionized before they are accelerated. Furthe~o~, the temperature of the Ha flare region is estimated as ~10,000 K on the solar disk (e.g. Svestka, 1969; de Jager, 1969). This suggests that the heavy ions are not fully ionized in this region, either. Therefore, we may say that, in the Ha flare and acceleration regions, the heavy atoms like Fe and Ni are only partially ionized. This means that, if we could have observed the bared heavy nuclei in solar cosmic rays, we should find the mechanism producing these particles in the acceleration region or its vicinity. 3. DE~LaP~~ OF SOLAR FLARES AND ITS RELATION TO TJBZ ACC~ERATION OF HEAVY NUCLEI

Solar flares are initiated when some instability is triggered in sunspot magnetic fields above active

105

2x105

4xm5

2X106

lOa T~PE~TURE

THE fONlZATlON

regions. This initial stage of solar flare development may be identified as the “flash” phase, which continues for lo-20 sec. It seems likely that this instability produces instantaneous acceleration of particles in the region where a flare is triggered. Although we do not know yet how long this acceleration continues, it is thought that both microwave impulsive and hard X-ray bursts are generated by energetic electrons (10-100 kev) produced by this instantaneous acceleration (e.g. Frost, 1969; Kane, 1973). As is well known, type III radio bursts are also often associated with these bursts. This acceleration seems to be independent of the development of the explosive phase, though, of course, considered as the precursor of this phase. This phase seems to be initiated by the generation of shock waves in the flare region (see, e.g. Wild et al., 1963). These waves would be produced as a result of the development of the instability mentioned above. In fact, shock waves are formed in association with the beginning of the explosive phase (see, e.g. Wild eb al., 1963; Sakurai, 1971b; Svestka, 1970). It seems likely that solar cosmic ray protons and heavier nuclei are mainly secondary accelerated during the explosive phase (Ellison et al., 1961; Svestka, 1970; Sakurai, 196Sa, 1971b). Therefore, we say that there must exist two steps for particle acceleration; i.e. the instantaneous and the “setondary” acceleration. The durations for these two accelerations seems to be less than 10-20 set and 100-300 set, respectively. These results necessarily require that the acceleration efficiency is very high for these two accelerations. As shown above, energetic electrons seem to be produced by the instantaneous acceleration. Since this acceleration also seems to be effective for protons and other nuclear particies, it may be possible to assume that these particles are accelerated to the relatively low energy of up to -10 MeV/nucleon,

42406

id'

&Id.

Te (K)

~UiliBRlU~

QF Fe 1ONS

FIG. 1. THE IONIZATION EQUILIBRIUM OF Fe IONS (JORDAN, 1969). THIS SITUATION SEEM ORSERVED IN THBREGIONWHEREASOLARFLAREOCCURS.

TO BE

The acceleration of heavy nuclei in solar llares although it does not seem that this acceleration produces particles of sub-relativistic or relativistic &rergies. Since the heavy nuclei of the iron group are not fully ionized, their acceleration efficiency is usually different from the particles fully ionized. At present, it is thought that Fermi acceleration is the most plausible explanation of the observed results on the rigidity spectra and the nuclear abundance of solar cosmic rays of energy 2 10 MeV/ nucleon (e.g. Hayakawa et al., 1964; Sakurai, 1965b,c; Wentzel, 1965). It seems likely that this acceleration mechanism also explains the acceleration of particles of the relatively low energy of up to 10 MeVlnucleon, sometimes to ~30 MeV/nucleon. If this is the case, for these low-energy particles, this mechanism is expressed as aR -=

at

A Moc2

y--g-~

where R, q, M,,, e, c and t are the particle rigidity, the coefficient of the Fermi acceleration, the rest mass of the particle, the electronic charge, the light speed and time, respectively. Here, A and Z are the mass and charge numbers of the particle, respectively. In the case in which particles are not fully ionized (see Fig. l), the charge number must be reduced according to the ionization state of the particles. In this case, we have to use the effective charge number Z*, which is smaller than Z. This means that the acceleration efficiency is higher for the particles with Z* than those with Z. Thus the charge states of the accelerated particles tend to produce an excess for the number of these species in comparison with the photospheric abundances. After acceleration in the flash phase, most of these accelerated particles up to -10 MeV/nucleon (sometimes, to -30 MeV/nucleon) seem to be ejected into outer space without farther acceleration. Such ejection may be accompanied by the passage of shock waves, generated in the flash phase, which, shortly after, become the source of type II radio bursts excited in the solar corona. While passing through the acceleration region immediately after generation, these shock waves seem to sweep away many of the accelerated particles. This action of the shock waves may be an important factor in determining the highest energy of these particles as -10 MeV/nucleon or so. It should be noted here that the beginning of the explosive phase is preceded by the generation of these shock waves in the case of proton flares (e.g. Sakurai, 1974b). Some function of the particles accelerated to a relatively low energy (say, 6 10 MeV/nucleon)

957

would remain without being swept away by shock waves when they are deep in the solar chromosphere or corona. These remaining particles would be fully ionized through their interaction with bombarding kev electrons simultaneously accelerated by the instantaneous acceleration during the flash phase, and then further accelerated to high energy (say, > 10 MeV/nucleon) by acceleration induced by the passage of shock waves (see, e.g. de Jager, 1969). This “secondary” acceleration takes place during the explosive phase, which is initiated with the generation of shock waves. In this case, the highest energy of accelerated particles is determined as that at which these particles can escape from the acceleration region, as given by R, in equation (3) (see, in detail, Sakurai, 1974b). The overabundance of the heavy nuclei in the energy 10 MeV/nucleon is thus necessarily produced by the following acceleration during the explosive phase. This “secondary” acceleration takes place during the explosive phase, and in this case, for medium and heavy nuclear particles, the ratio A/Z is about equal to 2. Therefore, the acceleration efficiency is almost the same for all these nuclear particles (e.g. Sakurai, 1965b,c). Since the ionization potentials for the electrons in the lowest energy states of these particles are given by: Q c 13*56Z2 eV,

(2)

the values of these potentials tend to increase with the charge number of particles. For examples, these values are given 1356 and 9156.6 eV for hydrogen and iron atoms, respectively. Because this ionization potential is very high for the iron-group atoms, it seems difficult for all of these atoms to be fully ionized by their interaction with keV electrons accelerated by the instantaneous acceleration mechanism. However, most of the medium nuclei would be completely fully ionized. The minor fraction of fully ionized heavy nuclei, which would be accelerated to energy 210 MeV/nucleon, would thus produce a slight deficit as seen in the abundance of the heavy nuclei of solar cosmic rays of this energy, though nearly the same as that of the photosphere of the Sun. Such a deficit could be seen in the results obtained by Fichtel and coworkers, though not conclusively (e.g. Fichtel, 1971; Biswas, 1972; Bertsch et al., 1972, 1973). As mentioned above, this deficit of heavy nuclei in the abundances of

958

IWNrT0M0 SAKURAX

solar cosmic rays seems to be highly dependent on the physical processes in the acceleration region, associated with the behaviour of accelerated particles. The model presented here suggests that the charge states of heavy nuclei and the ovembundan~ of the heavy nuclei in solar cosmic rays are both strongly dependent on the manner of the development of parent flares. In particular, the chemical composition of the particles of energy 210 MeV/ nucleon is considered as a result of their ionization by bomb~d~g keV electrons accelerated in the instantaneous process. Roughly speaking, the iron group abundances of solar cosmic rays of energy > 10 MeV/nucleon are similar to those of the photosphere, though rather lower (e.g. Bertsch et al., 1969, 1972, 1973). From this fact, we have concluded that the “secondary” acceleration during the explosive phase may be identified as the Fermi mechanism (see, e.g. Hayakawa et al., 1964; Sakurai, 1965b,c; Wentzel, 1965). Therefore, the overabundance in the heavy nuclei of energy ~10 MeV/nucleon may have been produced during the initial stage of a flare development, during which the partially ionized particles are mainly accelerated simultaneously with electrons of keV energy. The latter seems to become not only the source of microwave impulsive and hard initial X-ray bursts, but also works as an agent to ionize the former while they are further accelerated. 4. OVERABUNDANCE OF HEAVY NUCLEI IN SOLAR COSMIC RAYS

When the iron-group atoms are not fully ionized, the acceleration rate becomes charge-dependent because A/Z* is charge-dependent. Hence these atoms would be more easily accelerated to higher energies. If Equation (1) is applied, the final rigidity R, is given as:

where me and & are respectively the tenement time in the acceleration region and the initial rigidity of the particle (see, Sakurai, 1974b). In the above equation, we have replaced Z by Z*, because the atoms under consideration are partially ionized. The final rigidity obtained above is an important factor because the iterated rigidity spectrum of the accelerated particles is expressed as N(R) = G, exp (--RI%),

(4)

where K, is determined by the acceleration efficiency, the injection rate and the confinement time Q

(e.g. Roederer, 1964; Sakurai, 1974a,b). In principle, therefore, the excess of partially ionized particles with effective charge Z* is estimated by calculating the factor l& However, it does not seem easy to estimate which factor in &, is important in the production of the overabundance in the medium and heavy nuclei of low energy solar cosmic rays (& 10 MeV/nucleon). Because Z*
REFERENCES Armstrong,

T. P. and Krimigis, S. M. (1971). J. gee@.

Res. 76, 1230. Armstrong, T. P., Krimigis, S. M., Reames, D. V. and Fichtel, C. E. (1972). Lgeophys. Res. 77,3607.

The acceleration

of heavy nuclei in solar flares

Bertsch, D. L., Fichtel, C. E. and Reames, D. V. (1969). &rophys. J. 157, L53. Bertsch, D. L., Fichtei, C. E. and Reames, D. V. (1972). Astrophys. J. 171, 169. Bertsch; D. L., Fichtel, C. E., Pellerin, C. J. and Reames, D. V. (1973). Astroohvs. J. 180. 583. Biswas, S. (1972). In& .? Radio ipace Phys 1, 19. Biswas, S. and Fichtel, C. E. (1964). Astrophys. J. 139,941. BISWAS, S. and Fmrrra~, C. E. (1965). Space Sci. Rev. 4,709. Cartwright, B. C. and Mogro-Campero, A. (1972). Astrophys. J. 177, L43. Cartwright, B. G. and Mogro-Campero, A. (1973). High-Energy Phenomena on the Sun (Eds. R. Ramaty and R. G. Stone), p. 393. NASA, GSFC X-693-73-193. Cowsik, R. and Price, P. B. (1971). Physics Today, 24, No. 10, 30. de Jager, C. (1969). Solar Flares and Space Research (Eds. C. de Jager and Z. Svestka), p. 1. North-Holland, Amsterdam. Ellison, M. A., McKenna, S. M. P. and Reid, J. H. (1961). Dunsk Obs. Pub. 1, 53. Fichtel, C. E. (1971). Phil. Trans. R. Sot. London A270, 167. Frost, K. J. (1969). Astrophys. J. 158, L159. Hayakawa, S., Nishimura, J., Obayashi, H. and Sato, H. 1964). Prog. Theor. Phys. Suppl. 30,86. Hovestadt, D. (1973). Proc. 13th Znt. Conf. Cosmic Rays 5,3685. Jordan, C. (1969). Mon. Not. Roy. Astron. Sot. 142,501.

959

Jordan, C. (1970). Mon. Not. R. astr. Sot. 148,17. Kane, S. R. (1973). ZAU Symp. 57, Coronal Disturbances, Australia. Kane, S. R. and Lin, R. P. (1971). Solar Phys. 23,457. Mogro-Campero, A. and Simpson, J. A. (1972a). Astrophys. J. 171, L5. Mogro-Campero, A. and Simpson, J. A. (1972b). Astrophys. J. 177, L37. Price, P. B. (1973). Space Sci. Rev. 15, 69. Price. P. B., Hutcheon. I.. Cowsik. R. and Barber. D. J. (1971). ihys. Rev. iett: 26,916: Roederer, J. G. (1964). Space Sci. Rev. 3,487. Sakurai, K. (1965a). Rep. Zonosph. Space Res. Japan 19,405. Sakurai, K. (1965b). Pub. astr. Sot. Japan 17,403. Sakurai, K. (1965~). Rep. Zonosph. Space Res. Japan 19,421. Sakurai, K. (1971a). Pub. astr. Sot. Pacific 83,23. Sakurai, K. (1971b). Solar Phys. 20,147. Sakurai, K. (1974a). Astrophys. Space Sci. 28, 375. Sakurai, K. (1974b). Physics of Solar Cosmic Rays, pp. 458. University of Tokyo Press, Tokyo. Svestka, Z. (1966). Space Sci. Rev. 5,388. Svestka, Z. (1969). Solar Flares and Space Research (Eds. C. de Jager and Z. Svestka), p. 16. NorthHolland, Amsterdam. Svestka, Z. (1970). Solar Phys. 13,471. Wentzel, D. G. (1965). J.geophys. Res. 70,2716. Wild, J. P., Smerd, S. F. and Weiss, A. A. (1963). Ann. Rev. Astron. Astrophys. 1,291.